Identifiers used by Topaz and their Formats
DOI's
CrossRef acts as the registration agency for the PLoS DOI's. URL's (and other meta-data) for DOI's are registered with CrossRef, who in turn publish the URL's at CNRI's handle server.
http://dx.doi.org/ is an HTTP gateway to the handles system: any (registered) DOI can be resolved to it's (first) URL via an HTTP GET to http://dx.doi.org/<doi> - this will return a 302 with the URL in the Location header.
Another gateway is located at http://hdl.handle.net/ with the same usage to resolve DOI's: http://hdl.handle.net/<doi>
Parsing
(see DOI Numbering).
Syntax: <prefix>/<suffix>
Both prefix and suffix and case insensitive, but the canonical form is upper-case. Any printable character from UCS-2 is allowed in both prefix and suffix, except for '.' and '/' in the prefix.
Info URI's
General Overview, and RFC.
DOI's: info:doi/<doi>
Parsing
Syntax: info:<namespace>/<identifier>
The namespace is case-insensitive (SHOULD be lower case). The info scheme supports hierarchical processing, but can't in general be parsed according to the generic rules. The identifier part is parsed according to the rules for that namespace.
DOI URI's
The DOI site talks about doi:<doi> everywhere, and they provide plugins for Netscrape and IExploder to handle this. However, attempts to formally register the doi scheme seem to have failed (the internet draft is dead); similarly the URN:doi seems to be going nowhere. Currently the only official registration seems to be the info:doi/<doi> stuff.
Some more discussion on this can be found on DOI's Identifier Factsheet.
